|Home | About | Journals | Submit | Contact Us | Français|
We aimed to identify risk factors for recurrent syncope in children and adolescents with congenital long QT syndrome (LQTS).
Data regarding risk assessment in LQTS after the occurrence of first syncope are limited.
The Prentice-Williams-Peterson conditional gap time model was utilized to identify risk factors for recurrent syncope from birth through age 20 years among 1648 patients from the International LQTS Registry.
Multivariate analysis demonstrated that QTc duration (≥500 msec) was a significant predictor of a first syncope (HR=2.16), whereas QTc effect was attenuated when the endpoints of second-, third-, and fourth- syncope were evaluated (HR = 1.29, 0.99, 0.90, respectively; p<0.001 for the null hypothesis that all four HRs are identical). A genotype-specific sub-analysis showed that during childhood (0–12 years) LQT1 males had the highest rate of first syncope (p=0.001), but exhibited similar rates of subsequent events as other genotype-gender subsets (p=0.63). In contrast, in the age-range of 13–20 years, LQT2 females experienced the highest rate of both first and subsequent events (p<0.001 and p=0.01, respectively). Patients who experienced ≥1 episodes of syncope had a 6–12-fold (p<0.001 for all) increase in the risk of subsequent fatal/near-fatal events independently of QTc. Beta-blocker therapy was associated with a significant reduction in the risk of recurrent syncope and subsequent fatal/near-fatal events.
Children and adolescents who present following an episode of syncope should be considered to be at a high a risk for the development of subsequent syncopal episodes and fatal/near-fatal events regardless of QTc duration.
Congenital long QT Syndrome (LQTS) is an inherited cardiac arrhythmic disorder due to mutations affecting the cardiac ion channels resulting in delayed ventricular repolarization and a prolonged QT interval on electrocardiogram (ECG). The associated arrhythmias may manifest as syncope, aborted cardiac arrest, or sudden cardiac death (SCD).
Thus far, studies on the risk stratification of LQTS have provided data on the clinical and genetic risk factors predicting the likelihood of a first cardiac event (1–3). However, these studies may not apply to patients with LQTS who present after their first syncopal episode. Specifically, children and adolescents with congenital LQTS were shown to have a high-risk for first cardiac events. However, currently there are no data regarding risk factors for recurrent syncopal episodes in this population. We hypothesized that the traditional clinical, electrocardiographic, and genetic risk factors for a first syncopal episode would exhibit a time-dependent change in predictive value for recurrent syncope. Accordingly, the primary aim of the present study was to investigate risk factors for recurrent syncope in children and adolescents who present for evaluation following a syncopal episode. To further stress the importance of recurring syncope as a risk factor in LQTS, we also assessed the risk for the development of fatal or near-fatal events following repeated episodes of syncope in this population.
The study population was drawn from the International LQTS Registry. Patients were included in the study if they had 1) a corrected QT interval ([QTc] assessed using Bazett’s formulae ) ≥450 msec, and/or 2) a documented LQTS-causative mutation by genetic testing. Subjects were excluded from the analysis if they 1) had a QTc <450 msec on the baseline ECG without a genotype positive mutation; 2) were from families with known LQTS mutations but did not have their family’s LQTS mutation (the exclusion of such individuals might involve rare patients with an undetected second mutation); and 3) were enrolled in the registry after 20 years of age, to enable complete data collection regarding events that occurred during the childhood and adolescence periods. The final sample comprised 1,648 patients.
On enrollment, a complete past medical history was obtained, and patients were followed up yearly for additional medical information. Prospectively designed forms were used to obtain clinical data from patients, their families, and physicians (by regular phone and mail contact), regarding individual and family history, demographic, ECG, therapeutic and cardiac event information. On enrollment in the International LQTS Registry, a 12-lead ECG was obtained from each subject. Lead II was used for measuring QT interval, which was adjusted for heart rate by Bazett’s formula (4). Information on LQTS therapies was gathered prospectively at yearly intervals. Follow-up data regarding beta-blocker therapy included the starting date, type of beta-blocker, and discontinuation date in case it occurred. Among subjects who died, the usage of a beta-blocker before death was determined retrospectively.
The LQTS genotype was determined by multiple established genetic laboratories using standard mutational analytic techniques. In this study sample, 738 patients were genetically tested and identified to have an LQTS-causing mutation. The LQTS Registry study was approved by the University of Rochester Institutional Review Board, and informed consent was obtained from study participants or their guardians. The first recorded ECG obtained at the time of patient enrollment in the Registry was used in the current analysis.
The primary endpoint was the occurrence of recurrent syncopal episodes following the first event (categorized as 2nd through 4th episodes) from birth through age 20 years. Syncope was defined as a transient loss of consciousness with abrupt onset and offset. A subsequent syncopal episode was defined as syncope occurring seven or more days from the prior syncopal episode. The seven-day separation period was included to distinguish one event from a continuing event associated with specific triggers (i.e. fever, menses, medications), allowing patients to be clearly removed from the triggered event and stabilized on new therapies. Aborted cardiac arrest (ACA) and LQTS-related SCD served as endpoints for our analysis on the association between recurrent syncope and the risk of subsequent fatal or near-fatal events. ACA was defined as cardiac arrest with recovery after defibrillation, and SCD was defined as sudden and expected death due to LQTS.
The clinical characteristics of the study patients were categorized by the total number of syncopal episodes experienced by the patients. Patients with different numbers of syncopal episodes were compared using the chi-square test for categorical variables, and the ANOVA test for continuous variables.
The conditional model, proposed by Prentice, Williams, and Peterson (PWP), was used to simultaneously model the distribution of the time to first syncope and the gap times between subsequent syncopal events as functions of risk factors of interest (5), with inference based on the robust sandwich estimator of the covariance matrix for the estimated model parameters in order to account for potential dependencies between the gap times. Unlike some other models for recurrent events, such as the marginal model of Wei, Lin, and Weissfeld (6), the PWP model assumes that risk for a second event is conditional upon having a first event. Each ordered outcome (1st syncope, 2nd syncope, 3rd syncope, 4th syncope) is assigned to a separate time-dependent stratum, with its own baseline hazard function; and all covariates are interacted with each stratum, allowing each ordered outcome to effectively have its own unconstrained Cox model. Pre-specified covariates that were included as potential predictors of a first and recurrent syncopal episodes included: gender, age>13 years (except for first syncope, where this main effect term was absorbed by the baseline hazard function, since age is the time scale), gender x age>13 interaction (to allow for separate hazard ratios for males vs. females among those <13 and those 13+ years of age ), the Bazett-corrected QT interval (dichotomized at QTc ≥ 500 msec in the primary analysis, and at the upper quartile [≥540 msec] in a secondary analysis), and beta-blocker therapy (modeled as a time-dependent binary covariate).
We also carried out a genotype specific subset analysis among the 644 LQT1 or LQT2 study patients to identify potential genotype-gender factors as predictors of first and subsequent episodes of syncope. We focused on the LQT1 and LQT2 genotypes due to the limited number of patients with the other genotypes. Since previous studies have shown a relationship between genotype, gender, and age (1), we performed Kaplan-Meier estimates of time to a first and second event based on genotype-gender subsets (categorized as LQT1 males, LQT1 females, LQT2 males, and LQT2 females) and within pre-specified age-groups (childhood: 0 through 12 years; adolescence: 13 through 20 years), with the log-rank test used for hypothesis testing. Multivariate analyses of these interactions were performed to analyze genotype and gender interactions within age groups. Due to a lower sample size of genotyped patients, we limited the analyses in the genotype-specific models to a comparison between risk factors for the first- and the second- syncope (i.e. all subsequent episodes were categorized into a single ≥2 events endpoint).
Wald tests were used for each joint hypothesis test which determined whether the covariate by event strata interactions were significant.
The Cox model was carried out to evaluate the association between first- and recurrent- syncopal episodes and the risk of subsequent ACA or SCD. The same covariates used to model recurrent syncope in the PWP model were utilized in this analysis, with the addition of the time-dependent number of prior syncopal events. The Kaplan-Meier estimator was used to assess the 5-year cumulative probability of an ACA or SCD based on combined assessment of the number of syncopal episodes experienced by an individual and QTc duration.
Statistical analysis was performed using SAS 9.1.3 for Windows (SAS Institute, Cary, NC, USA)
The clinical characteristics of the total study cohort, and of the genotype-positive patient subset, by the number of episodes of syncope during follow-up are shown in Table 1A and B, respectively. Comparison of ECG parameters showed increasing QTc with increasing number of syncope. Notably, patients who experienced syncope had a very high frequency of beta-blocker usage (≥88%). In addition, the frequency of ICD usage was significantly higher among patients who experienced syncope, but similar among the 4 syncope subgroups (Table 1A). Among the 738 genotype-positive patients, the LQT1 and LQT2 genotypes comprised the majority (87%) identified LQTS genotypes (Table 1B). The mean ages at which the first through fourth episodes of syncope occurred were 9, 10, 11, and 12 years (± 6 years for all), respectively, and the majority (>80%) of subsequent events occurred within 5 years of the prior event. The frequency of patients who experienced ≥1 syncopal events was similar between LQTS patients who did- and did not- undergo genetic testing (37% and 42%, respectively).
Patients with QTc ≥ 500 msec had a significantly higher cumulative probability of experiencing their first syncope as compared to patients with QTc < 500 msec (75% vs. 48%, respectively; p = <0.001) from birth through age 20 years (Fig. 1A). However, among patients who experienced a first syncope, the rate of subsequent events was increased among those who had both QTc < 500 msec and those with QTc ≥ 500 msec (74% and 76%, respectively; p = 0.11) at age 20 years (Fig. 1B). Similarly, among patients who experienced 2, 3, and 4 syncopal episodes, the risk for subsequent syncope was virtually identical between patients who had narrow or prolonged Qc (not shown). This was supported by the PWP model for recurrent events (Table 2), which showed that QTc ≥500 msec was a powerful independent predictor (>2-fold risk increase; p<0.001) of a first syncopal episode; whereas the risk associated with QTc was significantly attenuated when the endpoints of subsequent episodes of syncope were assessed (2nd syncope: HR= 1.29; 3rd syncope: HR = 0.99, 4th syncope: HR = 0.90; interaction p-value < 0.001 for the null hypothesis that all four HR are identical). The results regarding the association between QTc and the risk of first- and recurrent- events were virtually identical when patients with congenital deafness (n=48) and multiple mutations (n=17) were excluded from the analysis. Furthermore, a secondary analysis, in which QTc was categorized at the upper quartile of the study population (≥540 msec [n=425]) or as a continuous measure, consistently demonstrated that QTc was predictive only of a first event and not of subsequent events (not shown).
A gender-genotype specific analysis, that was carried out in a subset of 644 patients having either the LQT1 (n=387) or LQT2 (n=257) genotypes, showed age-dependent risk factors for first- and subsequent syncope (Fig. 2 and and3).3). During childhood, LQT1 males had the highest cumulative probability for a first syncope (36% during 10 years of follow-up [Fig. 2A]). In contrast, after the occurrence of a first event, the cumulative probability of a subsequent syncopal episode during childhood was increased substantially in all genotype-gender subgroups (35–46% during only 2-years of follow-up [Fig. 2B]). Consistent with these findings, multivariate analysis showed that during childhood, LQT1 males exhibited a significant increase in the risk of a first syncope (p = 0.001 for the overall difference among the subgroups) but did not show a statistically significant increase in the risk of subsequent events (p = 0.34 for the overall difference among the subgroups).
After the onset of adolescence there was an age-gender risk-reversal, and LQT2 females showed the highest rate of both first a syncope (38% during 5-years of follow-up [Fig. 3A]), and subsequent events (58% during only 2-years of follow-up [Fig. 3B). Similarly, multivariate analysis showed that females with the LQT2 genotype in the age-range of 13 through 20 years exhibited a higher risk for both a first- and subsequent events as compared with the other gender-genotype groups (p = 0.001 and p = 0.02, respectively, for the overall difference).
After adjustment for QTc, gender, and time-dependent beta-blocker therapy, the occurrence of recurring episodes of syncope was associated with a pronounced increase in the risk of subsequent ACA or SCD (Table 3). As compared with no events, having 1 or 2 syncopal episodes were associated with >6-fold (p<0.001) increased risk of subsequent ACA or SCD; and having >2 episodes was associated with a >12-fold (p<0.001) relative risk of subsequent fatal or near-fatal events (Table 4). Combined assessment of the number of syncopal episodes and QTc (Fig. 4) showed that the 5-year cumulative probability of ACA or SCD was lowest among those who did not experience syncope (1% and 5% in patients with QTc < 500 msec and ≥ 500 msec, respectively); intermediate among those who experienced 1 or 2 syncopal episodes (11–14%); and highest among those who experienced > 2 syncopal episodes 17–21%). Notably, the effect of QTc on the risk of subsequent ACA or SCD was pronounced prior to the occurrence of a first syncope, and attenuated following the occurrence of syncope (Fig. 4).
Beta-blocker usage was associated independently with a significant and a substantial reduction in the risk of both first and recurrent syncopal episodes (Table 2). The magnitude of risk reduction associated with beta-blocker therapy was similar in LQT1 and LQT2 patients. Furthermore, treatment with beta-blockers was associated with a significant >70% reduction in the risk of subsequent fatal or near-fatal events among patients who experienced first- and recurrent- episodes of syncope (Table 3).
Previous studies on risk stratification in LQTS have been limited to identifying markers that predict the risk of a first cardiac event. Our study is the first to assess traditional risk markers for cardiac events in the LQTS populations as independent predictors for subsequent syncope. Unlike previous studies, we assessed recurrent syncope as an endpoint rather than a risk factor. We have shown that, in children and adolescents with congenital LQTS, there are important time-dependent changes in risk factors after a first syncopal episode. Importantly, the present study shows that the risk for subsequent syncopal episodes following the occurrence of a first event is increased regardless of QTc duration, despite the high usage of beta-blocker therapy (>90%) during this time-period. Furthermore, our results confirm that recurrent syncope is a powerful predictor of subsequent fatal or near-fatal events, independently of QTc. These findings have important clinical and therapeutic implications for LQTS patients who present for evaluation after the occurrence of a first episode of syncope prior to age 20 years.
The heart rate corrected QT interval has been the basis for diagnostic criteria for LQTS (7), and is a known predictor of a first cardiac event in this population (3,8–11). Our results are consistent with previous observations, demonstrating that a QTc ≥500 msec is an independent predictor of a first cardiac event. However, an LQTS patient with who has already experienced a syncopal episode should be considered at higher risk for a recurrent syncopal episode and possibly even LQT-triggered ACA/SCD regardless of his/her QTc value. Notably, the proportion of patients with a lower-range QTc in the total population (65%) was higher than among patients who experienced recurrent events (approximately 50%). However, while in the former subgroup a QTc < 500 msec is indicative of relatively lower risk for a first cardiac event (3,8–11), it is not a reassuring value after the occurrence of a LQTS cardiac event. It is possible that additional risk factors may be present in LQTS patients who present with a first syncope despite a lower-range QTc, including environmental factors, the presence of modifier genes, or high-risk ion channel properties of the LQTS-related mutation. Thus, risk assessment and management of clinically symptomatic children and adolescents with congenital LQTS should be independent of QTc. This impact on risk stratification requires careful evaluation to distinguish whether that first event stemmed from LQT’s signature dysrhythmia of torsades or whether that fainting episode was simply a vasovagally-mediated episode occurring in a LQTS host.
The effect of gender, genotype, and age on time to the first cardiac event has been well studied (1,7–9). During childhood, LQTS males were shown to have an increased risk for a first cardiac event compared to their female counterparts (12), and to experience the events at a younger age (1). After the onset of adolescence, however, risk-reversal occurs, and females were shown to exhibit a 2–3-fold increased risk of a first cardiac event as compared with males throughout adulthood (10,12). Prior studies have also shown that there are also important genotype-gender interactions within the age-groups. Thus, during childhood, LQT1 males were shown to be at higher risk for a first cardiac event, whereas after the onset of adolescence, LQT2 females were shown to exhibit a higher risk (1,12). Consistent with prior data, our genotype-specific analysis also shows that LQT1 males and LQT2 females experienced the highest rate of first events during childhood and after the onset of adolescence, respectively. However, our findings extend prior reports, and demonstrate that after the occurrence of a syncope recurrent event rates are markedly increased during childhood regardless of genotype or gender (≥35% in all genotype-gender during only 2 years of follow-up), whereas the risk for recurrent events during the post-adolescence period remains higher among LQT2 females. Notably, women with the LQT2 genotype who experienced a first syncope exhibited an extremely high rate of subsequent events (58% during only 2-years of follow-up), further stressing the importance of careful follow-up and timely therapeutic intervention in this high-risk population.
Syncope, although non-lethal, is associated with comorbidities such as trauma, and is a much more frequent cardiac event than ACA or SCD. The present study supports prior data regarding the importance of first- and recurrent- syncope as a powerful predictor of subsequent fatal or near-fatal events in LQTS children and adolescents (8–9). We have shown that after the occurrence of a first syncopal episode LQTS patients exhibited a 6-fold increase in the risk of subsequent ACA or SCD, and the risk of fatal or near-fatal events was increased to >12-fold among patients who experienced 3 or more episodes of syncope. These findings were independent of QTc or age-gender interactions. Furthermore, patients who had 3 or 4 episodes of syncope prior to age 20 years (who comprised approximately one fifth of the total study population and 43% of symptomatic patients) experienced a very high rate of subsequent ACA or SCD during only 5 years of follow-up (in the range of 17–21%, regardless of QTc). Thus, the occurrence of recurrent non-fatal events has important prognostic implications in LQTS patients, that are independent of the traditional risk factors previously reported in this population (1–3,7).
The mainstay of LQTS treatment in arrhythmia prevention is beta-blocker therapy, which suppresses the adrenergic-mediated triggers (13). Beta-blockers reduce cardiac event rates in LQT1 and LQT2 patients in a risk-dependent fashion (14). Beta blockers were shown to reduce lethal events by 70% in high risk children (8), and by 64% in high-risk adolescents with recent syncope (9). However, prior studies have also suggested that beta-blocker therapy may have important limitations in LQTS patients, including a relatively high-residual rate of cardiac events in patients with LQT2 and LQT3 (15) and a high residual rate of ACA or SCD among those who experience syncope during beta-blocker therapy (16). Our study confirms that beta-blocker therapy is associated with a significant reduction in the risk of a first cardiac event in LQTS children and adolescents (8.9). Furthermore, our findings extend prior data and demonstrate that treatment with beta-blockers is associated with a pronounced (>70%) reduction in the risk of subsequent ACA or SCD among patients who experienced any number of prior syncopal episodes. However, we have shown that the cumulative probability of ACA or SCD among patients who experienced >2 syncopal episodes was approximately 20% at 5-years of follow-up, despite the fact that the majority of this population was treated with beta-blockers. Thus, based on the findings of the present study and prior reports, a management strategy among LQTS patients who present for risk assessment following a syncopal episode may include the following: 1) initiation of beta-blocker therapy in patients who experience a syncopal event without therapy (17); 2) additional interventions, including primary ICD therapy (18) and/or LCSD (19–21) in patients who experience syncope during beta-blocker therapy; 3) careful follow-up for residual symptoms or arrhythmias after the initiation of beta-blocker therapy, with consideration of ICD therapy for patients who experience >2 episodes of syncope during follow-up (who received a similar frequency of ICD implants as patients who experienced 1 or 2 episodes, but had a very high rate of subsequent of ACA or SCD at 5-years [approximately 20%]).
In analyzing recurrent syncope, we censored ACA and SCD in the cause-specific PWP model, which may have biased subsequent risk group analyses. We addressed this bias with a secondary analysis on predictors for ACA or SCD, ensuring consistency in the risk analysis.
Data for the present study were derived from patients enrolled in the International LQTS Registry. Thus, similar to prior registry studies, it is possible that there were some inaccuracies in the yearly data collection and an inconsistent follow-up of some study patients. Furthermore, despite the fact that cardiac events were reviewed carefully by the study specialists, confounding factors, including vasovagal, neurologic, or metabolic disturbances, cannot be ruled out as the mechanism that precipitated some of the syncopal events in the present study.
It should be stressed that present findings regarding the lack of association between QTc duration and the risk for subsequent events are based on categorized data in a large study population. Thus, individual clinical assessment for therapeutic intervention is still important in this population (i.e. an extremely prolonged QTc in a symptomatic individual may still be considered as an additional marker for more aggressive intervention). Furthermore, it should be noted that the implications regarding risk factors for recurrent events that were identified in the present study are restricted to children and adolescents who experience a first syncope
Our sample population had a more limited number of genotyped patients (n=738), of whom the majority comprised the LQT1 and LQT2-genotypes. Therefore, risk factors for recurrent events were not assessed for study patients with the less frequent LQT3–8 genotypes. Furthermore, due to a lower sample size, we limited the analyses in the genotype-specific models to a comparison between risk factors for the first and ≥2 cardiac events.
This study focuses on recurrent syncope in children and adolescents with LQTS. We have shown that the predictive value of traditional LQTS risk markers change in patients who experience a first non-fatal cardiac event. QTc was shown to be a major predictor of a first syncopal episode. However, the risk associated with this traditional ECG marker was attenuated among patients who experienced a first event. Furthermore, we have shown important age-related genotype and gender differences in the risk for recurrent events in LQT1 and LQT2 patients. These time-dependent clinical and genetic factors should be considered in risk assessment and the management of LQTS patients who present for evaluation after the occurrence of a first episode of syncope.
This work was supported in part by research Grants HL-33843 and HL-51618 from the National Institutes of Health, Bethesda, Maryland, and by an unrestricted grant from BioReference Laboratories, Inc., Elmwood Park, New Jersey
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.